US 6992604 B2
Provided is a system and method for converting digital data audio data audio data that has a predetermined input sample rate, into an analog data signal. A system includes a digital to analog converter (DAC) including a digital processing portion configured to receive as an input the digital audio data and timing information, the timing information being representative of a time base of the input sample rate. The digital processing portion is similarly configured to digitally process the digital audio data and the timing information to produce serialized output data. The DAC also includes an analog processing portion configured to convert serialized data to an analog format. The digital processing portion operates in accordance with at least one clock having a corresponding clock rate wherein the corresponding clock rate is unrelated to the input sample rate.
1. A digital to analog converter (DAC) configured to convert input digital audio data to analog data, the digital audio data having a predetermined input sample rate, the DAC comprising:
a digital processing portion configured to (i) receive as an input the digital audio data and timing information, the timing information being representative of a time base of the input sample rate and (ii) digitally process the digital audio data and the timing information to produce serialized output data; and
an analog processing portion configured to convert the serialized output data to an analog format;
wherein the digital processing portion operates in accordance with at least one system clock having a corresponding system clock rate; and
wherein the at least one corresponding system clock rate is independent of the input sample rate.
2. The DAC of
3. The DAC of
4. The DAC of
5. The DAC of
6. The DAC of
7. The DAC of
8. The DAC of
9. The DAC of
10. The DAC of
11. The DAC of
12. The DAC of
13. The DAC of
14. The DAC of
15. The DAC of
16. The DAC of
17. A method for converting digital audio data having a predetermined input sample rate to analog samples in a digital to analog converter (DAC), the method comprising:
receiving timing information representative of a time base of audio input data, the audio input data having a predetermined input sample rate;
producing an input sample rate value based upon the received timing information, the input sample rate value being representative of the predetermined input sample rate and a desirable output sample rate; and
rate converting the audio input data based upon a system clock, the rate converting producing output samples in accordance with the input sample rate value, the output samples being output at the desirable output sample rate;
wherein the input sample rate is independent of the system clock.
18. The method of
19. The method of
20. An apparatus configured for converting digital audio data having a predetermined input sample rate to analog samples in a digital to analog converter (DAC), the apparatus comprising:
means for receiving timing information representative of a time base of audio input data, the audio input data having a predetermined input sample rate;
means for producing an input sample rate value based upon the received timing information, the input sample rate value being representative of the predetermined input sample rate and a desirable output sample rate; and
means for rate converting the audio input data based upon a system clock, the rate converting producing output samples in accordance with the input sample rate value, the output samples being output at the desirable output sample rate;
wherein the input sample rate is independent of the system clock.
This application claims the benefit of U.S. Provisional Application No. 60/505,022, filed Sep. 24, 2003, entitled “Method and System for Converting Digital Samples to an Analog Signal,” which is incorporated by reference herein in its entirety.
1. Field of the Invention
The present invention relates generally to the conversion of digital data to analog data. More specifically, the present invention relates to the use of digital to analog converters (DACs) to convert digital data to analog signals in recovery systems such as set top cable boxes.
2. Related Art
Currently available set top cable boxes generally include one or more DACs. Each DAC typically includes several clocks to provide timing to various components within the DAC. More particularly, these clocks ensure that the output of audio and/or video data samples is output from the DAC at the same sample rate that data is input to the DAC. For example, the currently available set top cable boxes may include formats based upon moving picture experts group (MPEG) data transport. To ensure proper operation of these MPEG conforming set top cable boxes, resolving timing issues, such as the correlation between input and output sample rates, becomes absolutely critical. Matching input and output sample rates in DACs are desirable in order to prevent overflow or underflow of data within the DAC.
Resolving timing issues is much more difficult in more complex set top cable boxes. Complex set top cable boxes may decode several data streams simultaneously, requiring several different system clocks. For example, one set top box might support two televisions, with each television supporting different programming. Additionally, it is contemplated that each television may be able to support features such as picture-in-picture, such that each television can display two different programs simultaneously. In these examples, each separate data stream associated with each of the separate features will require its own local clock. More specifically, to ensure proper synchronization, each of these local clocks must operate at a clock rate directly related to the input sample rate of its associated data stream. Thus, a single DAC capable of accommodating multiple data streams and multiple set top box features, will require several system clocks.
As set top cable boxes and other audio/video systems that use DACs become even more complicated, miniaturizing DACs and other critical system components becomes a significant design challenge. Thus, from this perspective, a single DAC that requires multiple clocks can impose significant limitations and demands on DAC chip designs with miniaturization requirements.
What is needed therefore is a mechanism to reduce the number of clocks required for a single DAC to accommodate multiple data streams. The multiple data streams are created as a function of complex audio/video programming features that operate at different input sample rates. More specifically, what is needed is a method and system to enable a DAC to use a single clock capable of operating independent of the input sample rate.
Consistent with the principles of the present invention as embodied and broadly described herein, the present invention includes a DAC configured to convert input digital audio data that has a predetermined input sample rate, into an analog data signal. The DAC includes a digital processing portion configured to receive as an input the digital audio data and timing information. The timing information is representative of a time base of the input sample rate. The digital processing portion is also configured to digitally process the digital audio data and the timing information to produce serialized output data. The DAC also includes an analog processing portion configured to convert serialized data to an analog format. The digital processing portion operates in accordance with at least one clock having a corresponding clock rate wherein the corresponding clock rate is unrelated to the input sample rate.
In the present invention, a DAC is provided that includes a system clock that is capable of running, for example, off of a fixed clock or a free running oscillator. That is, the system clock is unrelated to the input sample rate. In one exemplary embodiment of the present invention, the DAC is capable of receiving linearly encoded digital data having a sample rate between 10 and 100 KHz and converting to serial digital signals having a 27 MHz bit rate. The ratio between the two rates need not be a fixed ratio. One feature of this process is that up-sampling performed within the DAC is preferably formed in stages. For example an incoming audio signal at a sample rate at 48 KHz, is up-converted to 843.75 KHz, then to 3.375 MHz, and finally to 27 MHz. The 27 MHz signals are then fed to an off-chip differential analog filter. In other embodiments, the differential analog filter can be an on-chip implementation.
In the present invention, the independence of the system clock and the input sample rate facilitates the formation of the digital processing portion of the DAC on a single integrated circuit (IC) circuit chip. A significant savings in IC chip space can be realized by performing up-conversion in multiple stages. Savings are also realized by eliminating the requirement that the system clock be related to the input sample rate. Additionally, in the present invention all of the processing within the digital processing portion remains in the digital domain all the way off-chip. In other words, there is no longer a requirement to include analog circuitry on the IC chip.
In this manner, analog circuitry can be kept quieter by isolating it from the digital circuitry. Additionally, the removal of the analog circuitry from the chip contributes to a reduction in the overall chip size. Thus, integration of the DAC into larger digital systems such as set top cable boxes becomes simpler and less costly.
Further features and advantages of the present invention as well as the structure and operation of various embodiments of the present invention, are described in detail below with reference to the accompanying drawings.
The accompanying drawings, which are incorporated in and constitute part of the specification, illustrate embodiments of the present invention and, together with the general description given above and the detailed description of the embodiments given below, serve to explain the principles of the invention. In the drawings:
The following detailed description of the present invention refers to the accompanying drawings that illustrate exemplary embodiments consistent with this invention. Other embodiments are possible, and modifications may be made to the embodiments within the spirit and scope of the invention. Therefore, the following detailed description is not meant to limit the invention. Rather, the scope of the invention is defined by the appended claims.
It would be apparent to one skilled in the art that the present invention, as described below, may be implemented in many different embodiments of hardware, software, firmware, and/or the entities illustrated in the drawings. Any actual software code with the specialized controlled hardware to implement the present invention is not limiting of the present invention. Thus, the operation and behavior of the present invention will be described with the understanding that modifications and variations of the embodiments are possible, given the level of detail presented herein.
As shown in
The ASRC 202 receives, as an input, the left and right audio data 106 output from the PCM input module 102, shown in
It's generally desirable to have a substantial match between the input and output sample rates of any DAC. Before reception by the PCM input module 102, the audio data samples 106 could have originated, for example, within an MPEG broadcasting station. The audio data 106 can be representative of different audio channels of a satellite television (TV) transmission to a customer's set top box. In this example, the audio data 106 is transmitted at a fixed data rate, such as the 48 KHz rate discussed above, and would be broadcast to all other users at this same fixed data rate.
In the above example of an MPEG transmission, all data formatting, such as the establishment of sample rates, occurs in an upstream MPEG head-end (not shown). A receiving end DAC, such as the DAC 104, must then be configured to match the incoming sample rate of incoming data, such as the audio data 106. A standard technique used to achieve this sample rate match is that the up-stream MPEG head-end will typically time-stamp the audio data stream. That is, the MPEG head-end will encode time references within audio data streams, such as the audio data stream 106, to more precisely convey the sample rate to the receiving DAC. Such a time reference can be used to determine whether the input audio data 106 is being transmitted at the 48 KHz, 44.1 KHz, or 32 KHz rate.
More specifically, the MPEG head-end will encode the audio data samples 106 with program clock references (PCRs), as a time reference for the transmission sample rate. At the same time, separate MPEG system time code (STC) pulses will be sent along a separate data path as an indication of the true time base of the sample rate of the audio data samples 106. Thus, in the exemplary embodiment of
As noted above, a significant limitation of conventional DACs is the requirement to have a separate clock to accommodate each separate input sample rate value received within the DAC. In these conventional DACs, STC pulses enable each of these separate and independent clocks to lock on to the corresponding separate and independent input sample rate. This process is a classic technique used in the art to help system clocks run a little faster or a little slower in association with the input sample rate of their corresponding input samples. Ultimately, this process assists the system clocks in matching the clock of the input samples at the MPEG head-end.
However, as the complexity of DAC based audio and video systems increases, these traditional techniques become less practical. For example, system clocks typically include components such as voltage controlled crystal oscillators (VCXO). The higher the number of separate input sample rates a given DAC is required to handle, the greater the number of VCXOs that will be required on the chip, greatly increasing the chip's cost. Secondly, synchronizing the timing between multiple VCXOs and multiple data streams, each having their own input sample rate, is more difficult to achieve technically.
The present invention, however, eliminates the need of multiple system clocks to accommodate multiple input sample rates. In the present invention, a sample rate conversion technique is used instead of having separate system clocks, each running at a different sample rate. In the present invention, a single free running clock, such as a crystal oscillator, can be used. More importantly, this single free running clock can run at a clock rate unrelated to any input sample rate. In other words, a DAC constructed in accordance with an embodiment of the present invention can receive input samples at any input sample rate, and at the same time, have the entire chip running off of a single fixed system clock.
In the present invention, the rate manager 200 receives the STC pulses 214 as a mechanism to determine the time base associated with the input audio data 106. The STC pulses 214 provide a reliable assessment of how fast or slow the input audio data samples 106 are running relative to the fixed system clock 212. Although any clock rate can be used, in the exemplary embodiment of
In the exemplary embodiment of
In the example of
In the example of
Each time the ASRC 202 needs to output a new 843.75 KHz sample, it reads the fractional value 201 to determine the relative position of the output sample. For example, the occurrence of the input sample 304 (within the rate manager 202) serves as a trigger for the ASRC 202 to correspondingly produce the output sample 310 and all subsequent output samples, including a final output sample 317, that occur before the next input sample 306 is received. Each of the 16 output samples, 310 through 317, is produced within the timing boundaries of the input samples 304 and 306. And the position of each output sample, for example 310–317, is precisely determined by a corresponding one of the fractional values 201. Stated another way, each of the corresponding fractional values 201 is calculated in order to position the output samples 310–317 relative to the input sample 304.
Although other formats can be used, in the exemplary embodiment of
The underlying idea between the interaction of the rate manager 200 and the ASRC 202 is that although the system clock 212 is a fixed clock, the data running through the digital portion 108 is the same as it would be if the system clock 212 was locked to the input sample rate. As a result, the digital portion 108 can use a free running clock that is inherently low jitter, totally independent from and unrelated to the input sample rate.
Although the rate manager 200 functionally operates off the system clock 212, it outputs pulses locked to the input sample rate. For example, the STC pulses 214 come in at a 27-MhZ rate. Samples 304 and 306, graphically illustrated in
The fractional value 201 enables the ASRC 202 to know that any given instant in time, how far through the input a given output sample will be. Stated another way, the fractional value 201 conveys how far each output sample is from the last input sample that was generated.
The rate manager 200 has two assessments of the input sample rate. First, it has a theoretical rate based on calculation of the fractional value 201. However, the rate manager 200 also knows the actual input sample rate based on the received STC pulses 214. With this information, the rate manager 200 is able to adjust the calculation of the output samples based on small changes in the input sample by comparing its calculation of where the input sample should be.
For example, as noted earlier the input samples are calculated to occur every 562.5 clock cycles, as indicated with input samples 304 and 306 shown in
After the ASRC 202 reads a sufficient number of fractional values 201 and correspondingly produces 16 consecutive output samples, as indicated in
As shown in
The structure of the polyphase filter 400, or any other suitable filter structure, allows much of the processing to be done at the input sample rate, with only a few multiply-accumulates at the output rate. Each individual 28-tap filter in the FIR filter engine 402 calculates a new Taylor coefficient each time a new input sample arrives. The total number of multiplies per input sample per channel is 224 (28 times 8).
Coefficients are passed to the fractional engine 404 using ping-pong buffers (not shown) in RAM. This allows the FIR filter engine 402 to take an entire input sample period to calculate new Taylor coefficients while the fractional engine 404 uses the previous coefficients.
The following equation is calculated at a 843.75 kHz output rate to create each output sample at the desirable 843.75 kHz rate discussed above. (In the equation x is a fraction, s is a scale factor, and y is the audio data to be sent to the CIC interpolator 206.)
This requires eight multiplies for each output sample-double that is required to calculate left and right values. The sample buffer (not shown) can be put in test tone mode and written by software for testing purposes.
Samples from the fractional engine 404 can be scaled before going to an RF modulator output 406, the peaking filter 204, and the CIC 206. This is meant to normalize gain through the polyphase filter 400. Scaling is a factor from 0 to 1, inclusive. Therefore, there is only attenuation, that should not exceed unity gain in the scaler.
When software changes the scale target, a counter will linearly ramp the actual scale factor from the current value to the new target value. The slope of the ramp is controlled by a register. The slowest ramp can take as long as 4.8 seconds to go from no attenuation to mute. The fastest ramp takes about one 843.75-kHz period.
The ASRC 202 uses a shared parallel multiplier to perform all multiplies in the FIR filter engine 402 and the fractional engine 404. The fractional engine 404 does 16 multiplies every 843.75 kHz for a stereo pair, plus 16 multiplies for the peaking filter 204, or 27 million multiplies per second total. The FIR filter engine 402 does 448 multiplies every input sample time, or 22 million multiplies per second total for 48 kHz.
The multiplier is timeshared, on a clock-by-clock basis, between the FIR engine 402 and the fractional engine 404. Each function gets access every other clock. Access to the RAM is also shared between the FIR filter engine 402 and the fractional engine 404. This means the minimum system clock frequency must be more than twice the fractional multiply rate or twice the FIR multiply rate, whichever is greater. For input sample rates 48 kHz and below, the fractional engine limits the clock rate to 54 MHz or higher. For input sample rates 96 kHz and above, the FIR engine 402 limits the clock rate to at least 448 times the sample rate.
Step 606 includes rate converting the audio input data based upon a system clock. The rate converting produces output samples in accordance with the input sample rate value, wherein the output samples are output at the desirable output sample rate. In the present invention, the input sample rate is unrelated to the system clock 212.
Referring back to
The ASRC 202 checks the PCM input 403 for zeros using a PCM zero detect module 408. If the zero detect module 408 detects a programmable number of zero samples in a row, it signals the digital modulator 209 to auto-mute its input. As soon as a non-zero PCM input value is detected, auto-mute will be de-asserted. By way of background, muting improves signal-to-noise measurements because the modulator 209 will not send constant zeros to the mapper 210 even when the PCM data is zero. However, muting causes the mapper 210 to output a square wave or constant low. The auto-mute can be enabled or disabled by software.
There is a pipeline delay through the ASRC 202, modulator 209, and the mapper 210. Since the detection is at the PCM input 403 to the FIR engine 402, and the mute occurs in the mapper 210. The mapper 210 mutes after dithered/modulated zeros have been output for some time. The mapper 210 un-mutes while zeros are still passing through the ASRC 202 and the modulator 209. Therefore, only zero PCM values will be auto-muted.
Although other modulation schemes can be used, the output of the modulator 209 is pulse width modulated. That is, non-zero level data appears at the output as a wide pulse data sample, and zero level data is intended to appear as a narrow square-wave pulse. The problem is that zero-level input data samples, received at the input to the modulator 209, do not always neatly appear as corresponding narrow square-wave pulses at the output of the mapper 210. This occurs because the zero-level data is actually modulated within the modulator 209, and subsequently can appear as “modulation noise” at the output. Muting the input to the mapper 210, however, produces cleaner narrow square-wave pulses at the output of the mapper 210.
When the muting 704 occurs, a cleaner signal is produced at the output of the analog processing portion 110 than would otherwise be produced without muting. Without muting, zero level data will propagate through the system, is then modulated within the modulator 209, and can be sensed as modulation noise at the output of the modulator 209. Muting, however, greatly reduces the occurrence of such modulation noise. When an end 706 of the consecutive zero level samples is sensed by the zero detect module 408, the input to the mapper 210 is un-muted. As indicated in step 808 of
Referring back to
In an associated external filter, discussed below with regard to
Each comb, for example combs 902 and 904, calculates a new output, which is the difference between the current and previous 843.75-kHz sample, every 843.75-kHz period. An integrator 906 adds a new value every 3.375-MHz period. A middle comb/integrator pair 908 with a zero stuffer 910 between is equivalent to a 0-order hold, or sample and hold (i.e., a bank of flip flops updated every 843.75 kHz). That is, the output of the first comb 902 can feed the input of a final integrator 912 directly as long as the comb holds its output constant for 32 27-MHz periods.
The entire CIC 206 can be bypassed, giving the effect of a first order CIC, which, as noted above, is the same as a 0-order hold. If a 1st-order CIC is sufficient, the CIC 206 could be removed entirely in a future design for some area reduction.
The present invention reduces the number of clocks required for a single DAC to accommodate multiple data streams. This feature is supported by provided a single free running clock that operates independent of the input sample rate. Since the requirement of multiple system clock to accommodate multiple data streams is eliminated, the digital processing portion 108 can be formed on a single IC, thus providing significant savings in terms of the cost and chip real estate.
The present invention has been described above with the aid of functional building blocks illustrating the performance of specified functions and relationships thereof. The boundaries of these functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternate boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed.
Any such alternate boundaries are thus within the scope and spirit of the claimed invention. One skilled in the art will recognize that these functional building blocks can be implemented by analog and/or digital circuits, discrete components, application-specific integrated circuits, firmware, processor executing appropriate software, and the like, or any combination thereof. Thus, the breadth and scope of the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.
The foregoing description of the specific embodiments will so fully reveal the general nature of the invention that others can, by applying knowledge within the skill of the art (including the contents of the references cited herein), readily modify and/or adapt for various applications such specific embodiments, without undue experimentation, without departing from the general concept of the present invention. Therefore, such adaptations and modifications are intended to be within the meaning and range of equivalents of the disclosed embodiments, based on the teaching and guidance presented herein. It is to be understood that the phraseology or terminology herein is for the purpose of description and not of limitation, such that the terminology or phraseology of the present specification is to be interpreted by the skilled artisan in light of the teachings and guidance presented herein, in combination with the knowledge of one of ordinary skill in the art.